search

influxdata / influxdb

Scalable datastore for metrics, events, and real-time analytics
https://influxdata.com
Apache License 2.0
28.29k stars 3.52k forks source link

Add support for influxdb.to to the 1.x line #17790

Open nathanielc opened 4 years ago

nathanielc commented 4 years ago

The 1.x line does not implement the influxdb.to() function as it is different from the 2.x line.

Clients using Flux inside the 1.x line need the ability to use influxdb.to() in order to downsample data using Flux.

nathanielc commented 4 years ago

We may want to hold off on this till #17791 is complete so as to not have to re-implement it when that change lands.

rbetts commented 4 years ago

Needs an acceptance criteria decision for fine grained auth in 1.x enterprise.

samhld commented 4 years ago

@rbetts @nathanielc In my experience, the need for FGA comes up rarely so it seems totally fine to leave that out, to me.

chobbs commented 4 years ago

Agree with @samhld - the FGA rarely comes up - were safe to exclude. @rbetts @nathanielc

timhallinflux commented 4 years ago

the 2.x write API exists now on 1.8, is that "helpful" in this regard?

samhld commented 4 years ago

@timhallinflux that's what we plan to use. The issue is specifically that there is no to() function so data can't be written back to the instance, so no downsampling can be done even when cron'ing interactive query. One would have to handle the response data in a client themselves and, in Canada Pacific Railway's case, they are going to be doing it in Bash given client restrictions and doing that in Bash is probably pretty painful.